Agreed. Apple have made the likelihood of gaming on Macs taking off almost zero. They're have to change so much, and throw a ton of money at developers too, as the market share is so tiny.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
M3 Pro Chip Barely Faster Than M2 Pro in Unverified Benchmark Result
- Thread starter MacRumors
- Start date
- Sort by reaction score
So if it ends up being accurate when others do tests, will you retract this?
That's the exact same iterative, non-innovational path that saw Intel losing to AMD and Apple making the pivot to the M1.
Explosive leaps in performance are what will be setting Apple ahead and continue to have it differentiate itself from other CPU makers.
100% agree. For my software (Cubase) I need maximum CPU for live audio processing, a fairly small amount of ram (16 or 32) and next to no GPU.
It’s pretty annoying now to have to pay for high end GPU to get high end CPU.
A maximum of 6P cores in a Pro chip is just lame given 2 generations ago there was already 8.
We have had one unverified test and like lemmings we get a chorus of, “The sky is falling! The sky is falling!”
Yeesh!
The better commentators on Youtube at least have the decency to say they have to see/run firsthand real world tests before making more definite conclusions.
Yeesh!
The better commentators on Youtube at least have the decency to say they have to see/run firsthand real world tests before making more definite conclusions.
Incidentally stores always hold a (very) few units of the most maxed out spec you can get.
Yep the M3Mx for an extra $1,499 AUD.
Amazing!!
Last edited:
I don't know if this link has already appeared on this very long thread but I think Snazzy Labs eloquently explains a lot of what we are seeing. It sounds like Apple were determined to get this new chip out to head off the ARM chips coming soon for Windows. In order to do this they had to go with the earliest versions of the 3nm despite the high rejection rates.
More here: -
More here: -
Not sure what you’re smoking?
Cinebench comparison:-
M3 +18% single core
M1Pro +18% multi core.
You can pick up a refurb 16” M1Pro 16/512 for $1839 for basically the same price as the M3 14” ($1,799). That’s a pretty good deal for the 16” given it’s latest price rise to $2,499 for the m3 pro.
Well not many M1Pro users will either.
I was hoping to get an 8P core M3 pro, not a 6P one. So I’ll wait till M4 Pro and see what pans out.
I was also planning to upgrade my old iMac i7, but with 16/512 configuration it’s $3,099 AUD which is pretty steep price to pay for a smaller screen ( and a step down from my 32/1TB configuration).
I’ll probably pick up a Mac Mini M2 16/512 when one comes up in the Australian refurb store in the future and pair it with a 27 or 32” screen and save a good $1000 or more. No doubt the Mac Mini M3 upgraded to 16/512 will be more expensive anyway for just a 15% speed bump anyways.
Thank you for doing the math. There is plenty of room—the reason to geld the chip is to drive sales to the Max.
Those numbers include services which have much higher margin than hardware.
The gross margin for hardware seems to be 35-36% based on Q4 numbers which is in the normal range for Apple.
My bad. The point I really wanted to make was that 8086/8088 only had a 16-bit instruction set (with nasty segmented addressing) whereas the 68k series had a true 32 bit instruction set (and a continuous 24-bit address space) from the get-go - even though the 68000 itself was physically 16 bit - so when the fully 32-bit 68020/30 came along the software didn't have to change, and (unlike x86) we didn't get abominations like near and far pointers in C.
Other processors were available, but the 68k would have made a far more future-proof basis for an "industry standard". Using 68k in the Mac and never having to accommodate legacy 8/16 bit code probably contributed to Apple's ability to later switch to PPC -> ia32 -> amd64 -> ARM64.
Last I heard, Intel was considering dropping 16 bit support, and Windows only fully dropped native 16 bit support with Windows 11.
As soon as they make a Mac Mini or iMac with M3 Max, I’m in. It better be FULL bandwidth too. Not the shenanigans they pulled with the M3 iMac only having 100GB versus the MBPs 150. (which yes is already lower than the M1/M2 version)
Last edited:
Today's headline in The Onion: "One Geekbench report noted by a sensationalist youtuber causes an avalanche of buyer's remorse..."
Last edited:
We saw decent improvements between the base M2 to M3 and the M2 Max to M3 Max. Clearly Apple went in a different direction with the M3 Pro. It has fewer transistors than the M2 Pro. The fact that it has any improvement in multicore despite 2 fewer performance cores is quite a feat.
You mean like Intel has done since forever? Intel has a turbo boost speed that can only be achieved with a single core. When Intel detects a single core benchmark it turns off all but one core then boosts the hell out of that core (6Ghz now).
Apple increased the number of efficiency cores and reduced bandwidth to increase efficiency, reduce heating and increase battery life. For most users, the M3 Pro has waaaaaaaaay more performance than will ever be used. Those uses simply will not have a better experience using their laptops by increasing the performance cores. However, they will have a better experience with a cooler, quieter laptop that has longer battery life. And for those who actually need more performance, Apple has an option for you too.
Here's my guess (might have been already mentioned here but I am not going to read 600+ messages): It seems that Apple realized that they have put too much capability into the basic M chip and are trying to adjust it by spreading the performance across the lineup within which one clearly see a progressive increase from the basic to Max and the Ultra, while there still is a discrepancy between chip generations. Basically, M1 and M2 were too impressive to their own detriment.